Systems and methods for photochemical reactions

Abstract

A photochemical reactor can include a plurality of flow cells whose illuminated internal surfaces coincide with a Φ = Φ * iso-flux boundary. Identical cells can be connected by short, small-diameter interconnects to form a cell chain of pitch p, and multiple chains can be arranged in parallel to meet throughput while maintaining residence-time and pressure-drop limits. Each cell can be illuminated by opposed, orthogonally oriented LEDs in 2-LED, 4-LED, or 6-LED topologies. In some embodiments, 6-LED illumination can be provided by a wireless cube LED apparatus having six emitters on orthogonal faces, a centrally located receiver coil, rectifier, and current driver. The cube can insert into keyed pockets that set optical standoff δ and orientation. A controller can sweep a transmitter frequency to select an operating frequency f_T* that maximizes received power for the installed receiver-coil geometry.

Description

Attorney Docket No.: 064752-501001 WOSYSTEMS AND METHODS FOR PHOTOCHEMICAL REACTIONSCROSS-REFERENCE TO RELATED APPLICATIONS

[0001] This application claims the benefit of U.S. Provisional Application No. 63 / 694,023, filed September 12, 2024, entitled “Systems and Methods for Photochemical Reactions,” the entirety of which is hereby incorporated by reference.FIELD

[0002] The present disclosure relates generally to photochemical reactors.BACKGROUND

[0003] Conventional plate and tubular photoreactors exhibit geometry-driven irradiance minima. To raise dark regions to a target dose, operators over-illuminate bright regions, increasing energy consumption, heat load, and reaction variability, which can be particularly detrimental in photoredox and advanced oxidation / reduction processes. Uniformity problems worsen with optical attenuation and with scale-up, while hydraulic limits (pressure drop, fouling risk) constrain channel dimensions.SUMMARY

[0004] Devices, apparatuses, systems, and methods for photochemical reactions are described.

[0005] One embodiment of a device for photochemical reactions is a flow-cell device for a photochemical reactor. The device can include a body defining an interior volume bounded by an inner wall. The inner wall can be shaped to substantially coincide with an iso-flux boundary at a target photon-flux threshold (0*). The iso-flux boundary can be the locus of points at which a photon-flux density, as evaluated by an attenuation model accounting for absorption (a) at an operating wavelength and emitter angular emission, equals <D* under a design illumination layout and an optical standoff distance (8) between emitters and the inner wall. The design illumination layout can be a two-emitter layout comprising two emitters as an opposed pair on a common axis, a four-emitter layout comprising four emitters as two opposed pairs on two mutually orthogonal axes, or a six-emitter layout comprising six emitters as three opposed pairs on three mutually orthogonal axes. The inner wall shape can be selected from a parametric family of shapes defined by a, O*, 8, and the selected layout.Attorney Docket No.: 064752-501001 WO

[0006] In some embodiments, the body includes emitter-reference features comprising planar standoff rails or keyed pockets positioned to locate light-emitting faces at the standoff 5.

[0007] In embodiments in which layout is the two-emitter layout, an emitter axis can define a symmetry axis, and the inner wall exhibits axial symmetry about that axis and mirror symmetry across a midplane normal to the symmetry axis. In embodiments in which layout is the four- emitter layout, two orthogonal emitter axes can define symmetry axes, and the inner wall can exhibit two orthogonal mirror planes and bilateral symmetry with respect to each axis. In embodiments in which layout is the six-emitter layout, three mutually orthogonal emitter axes can define symmetry axes, and the inner wall can exhibit triaxial symmetry with respect to those axes.

[0008] In some embodiments, a coincidence tolerance between the inner wall and the iso-flux boundary is a maximum normal offset less than or equal to 0.20 R eff. Emitters can be located outside the wetted volume, and the emitter-reference features can set 8 relative to exterior surfaces such that the iso -flux boundary coincides with the inner wall within the coincidence tolerance. In some embodiments, a coefficient of variation of photon-flux density over {<D > <D*} can be less than or equal to 1 under the design illumination layout at standoff 8. In some embodiments, a coefficient of variation of photon-flux density over {G > <D*} can be greater than 1 under the design illumination layout at standoff 8.

[0009] In some embodiments, the maximum normal offset between the inner wall and the iso-flux boundary can less than or equal to 0.10 R_eff. In some embodiments, the maximum normal offset between the inner wall and the iso-flux boundary is less than or equal to 0.05 R eff. In some embodiments, 0.1 R eff < 8 < 0.5 R_efif.

[0010] In some embodiments, the inner wall comprises smoothly joined surfaces whose minimum radius of curvature satisfies r min > 0.1 R eff.

[0011] In some embodiments, the parametric family of shapes is computed or generated by an optimization that minimizes a uniformity metric over { > *} while constraining 8 and the selected layout. The inner wall can be represented in manufacture by a superellipsoid (|x / a|Am + |y / b|Am + |z / c|Am)A(l / m)=l with parameters a,b,c,m chosen such that the maximum normal offset to the iso-flux boundary at * is < 0.20 R eff.Attorney Docket No.: 064752-501001 WO

[0012] D_min can be a minimum principal axis length of the flow cell, defined as the smallest principal diameter measured between opposing points on the inner wall along any illumination-defined principal axis. In some embodiments, the selected shape satisfies min(2a, 2b, 2c) > D_min, where a, b, c are principal half-spans along those axes.

[0013] In some embodiments, if a computed family member would yield a principal diameter < D_min, the selected member is the nearest member that meets D_min while maintaining the coincidence tolerance.

[0014] In some embodiments, the device is configured to mate with cylindrical interconnect capillaries of inner diameter d_c and length L_c satisfying d_c < 0.2 R_eff and L_c < 0.5 R_eff.

[0015] In some embodiments, emitter face centers lie on an emitter-reference surface S8 defined as the outer normal offset of the inner wall by 8, and emitter face normals are orthogonal to S8 within ±3°. Emitter face normal can deviate by less than or equal to 3° from the local normal of S3 and emitter-to-wall standoff can deviate from 8 by less than or equal to 0.10 R_eff.

[0016] In some embodiments, two emitter centers are colinear on the symmetry axis and approximately equidistant from a geometric center of the flow cell. In some embodiments, emitter centers form two opposed pairs on mutually orthogonal axes. Each center can lie on S8, and pairwise distances along each axis can be approximately equal.

[0017] In some embodiments, six emitter centers form an octahedral set on three mutually orthogonal axes, each center lies on S8, and all six center-to-center distances from a cell center are equal within ±10%.

[0018] The inner wall can carry a protective barrier coating that is chemically resistant and optically transmissive at the operating wavelength. In some embodiments, the barrier is selected from fluoropolymers, cyclic olefin copolymers (COC / COGS), polyacrylates / epoxies, PMMA, or polycarbonate.

[0019] In some embodiments, the device includes a photoreagent layer over the barrier configured to convert incident photons into reactive species. The photoreagent can be selected from TiO2, ZnO, sulfite-bearing chemistries, or indole-3 -acetic acid and can have a target thickness between 10 and 100 run.Attorney Docket No.: 064752-501001 WO

[0020] In some embodiments, the device includes an adsorbent porous layer overlying the barrier or the photoreagent to pre-concentrate reactants at the reactive interface. The adsorbent can be selected from silica gels, zeolites, or octadecylsilane-treated porous media.

[0021] In some embodiments, the inner wall carries a reversibly switchable dipolar surface layer configured to modulate adsorption / desorption for chromatography or capture / release operations, the layer including spiropyran / merocyanine or azobenzene chemistries operable by optical stimulus.

[0022] In some embodiments, one or more coatings carried by the inner wall can be applied by spray, dip, spin, sol-gel, interfacial polymerization, or layer-by-layer assembly. The resulting stack can maintain a coincidence tolerance of less than or equal to 0.20 R_eff.

[0023] In some embodiments, a dry-side reflective layer is disposed behind a transparent inner wall region to improve photon-flux uniformity. The reflective layer can be specular (protected Al or Ag) or dielectric stack or diffuse PTFE-class and arranged so that the coefficient of variation over {O > O*} is < 0.2.

[0024] In some embodiments, windows are bonded into flush pockets of the inner wall, the windows having anti-reflection coatings on an air side and sealed by welded frames or compatible adhesives, while not shifting the iso-flux boundary by more than 0.20 R_eff.

[0025] In some embodiments, emitters are inserted directly into keyed pockets that open to the wetted volume and are sealed by a transparent encapsulant or window. The optical plane of the encapsulant / window can be co-planar with the inner wall within the tolerance of less than or equal to 0.20 R eff, and 8 can be gauged to said optical plane. The exposed window or encapsulant face can carry one or more of the aforementioned coatings.

[0026] One embodiment of an apparatus for photochemical reactions can be a photochemical reactor apparatus that includes a headplate with an array of through-ports arranged on a lattice of lateral pitch (s) and a plurality of flow cell chains. Each chain can include a sequence of flow cells connected in series by cylindrical interconnect capillaries of inner diameter (d_c) and length (L_c).

[0027] In some embodiments, each flow cell comprises an inner wall shaped to substantially coincide with an iso-flux boundary at a target photon-flux threshold ( *) with a maximumAttorney Docket No.: 064752-501001 WO normal-offset tolerance < 0.20 R_eff. The iso-flux boundary can be evaluated under one of the following design illumination layouts: (i) two emitters as an opposed pair on a common axis, (ii) four emitters as two opposed pairs on two mutually orthogonal axes, or (iii) six emitters as three opposed pairs on three mutually orthogonal axes. An attenuation model can be used to account for absorption (a) at the operating wavelength and emitter angular emission, with an optical standoff distance (8) between emitters and the inner wall.

[0028] In some embodiments, inlet and outlet headers can be sealed to opposite faces of the headplate and formed as dished plenums in fluid communication with the through-ports.

[0029] The apparatus can include a plurality of LED emitters configured to illuminate flow cells of the sequence of flow cells. The plurality of LED emitters can be arranged according to a design illumination layout selected from (i) a two-emitter layout comprising two planar LED plates positioned on opposite sides of a flow-cell chain and registered by standoff spacers to set 8, (ii) a four-emitter layout comprising LED strips mounted on four faces of a rectangular LED column positioned between adjacent chains to present two opposed pairs around each neighboring cell, and (iii) a six-emitter layout comprising a unified chain in which cube-LED positions and flow-cell positions are mechanically integrated at a repetition interval pitch (p) and provided in flow-first and cube-first types installed in an alternating, laterally-staggered pattern so that each flow cell is surrounded by six cube-LED positions and each cube-LED position is surrounded by six flow-cell positions, each at standoff 8 within a maximum normal-offset tolerance < 0.20 R_eff. For each flow cell chain, the interconnect capillaries can satisfy d_c < 0.2 R_eff and L_c < 0.5 R_eff.

[0030] In some embodiments, headplate registration features include sockets or rails set the lateral pitch (s) and a stagger offset to maintain 8 within the maximum normal-offset tolerance across the bundle.

[0031] In some embodiments, for the two-emitter layout, each planar LED plate comprises a two-dimensional LED array on a thermally conductive substrate and includes manifolded microchannels or a serpentine liquid passage with inlet and outlet ports routed outside the process volume. In some embodiments, for the two-emitter layout, adjacent chains share back-to-back LED plates mounted to a common carrier that includes cooling passages or sealed cooling corridors hydraulically isolated from the process fluid. In some embodiments, for the two-emitter layout, each planar LED plate has an emitter line that is coincident with the chain axis within ALAttorney Docket No.: 064752-501001 WO< O.l R eff, and the plate normals are arranged symmetrically about a chain midplane within ±3°.

[0032] In some embodiments, for the four-emitter layout face-to-chain spacing equals 5 on each side so the iso-flux boundary remains coincident with the inner wall, and the columns and chains are arranged on a square checkerboard lattice such that each chain is bounded by four LED columns and each column serves four chains. In some embodiments, for the four-emitter layout, each LED column includes a thermally conductive core with cooling passages or sealed cooling corridors hydraulically isolated from the process fluid. In some embodiments, for the four-emitter layout, the LED-strip emitter centers are indexed to cell midplanes at the pitch (p) within ±0.2 p, and strip planes are parallel to a chain axis within ±3°. In some embodiments, for the four-emitter layout, each LED strip column has a column axis registered to the adjacent cellchain axis with AL < O.lO R eff and yaw / pitch < 3°.

[0033] In some embodiments, for the six-emitter layout, the unified chain is a wired chain and conductors for the cube-LEDs are routed along the chain to a feedthrough on a dry side. In some embodiments, for the six-emitter layout, the unified chain uses wireless cube-LED assemblies that are inductively powered by at least one transmitter coil, and the assemblies are positioned so that emitter face centers lie on an emitter-reference surface S3 defined as the outer normal offset of the inner wall by 8, with face normals orthogonal to S3 within ±3°, and spacing set to 8 within the maximum normal-offset tolerance. In some embodiments, centers of the six lateral cube-LED positions around each flow cell define a regular hexagon in plan with edgelength uniformity within ±10% across the bundle. In some embodiments, for the six-emitter layout, the six lateral cube-LED centers around each flow cell lie on a ring of radius R hex in plan view, with |R_hex - (8 + t_wall)| < 0.10-R_eff and edge-length uniformity within ±10%, where t_wall is a local wall thickness at a bulb equator.

[0034] In some embodiments, a chain axis is defined by centers of successive flow cells in a chain, and an LED array centerline of an emitter plate, strip column, or cube-LED spine is parallel to the chain axis within ±1° and has a maximum lateral offset Al < 0.20 R_eff relative to the chain axis along at least 80% of a chain length.

[0035] In some embodiments, a standoff uniformity satisfies |8(s) - 8| < 0.1 R eff along any emitter face spanning a bulb, within the maximum normal-offset tolerance. A bulb-to-neck geometry can satisfy a diameter ratio D_bulb / d_c > 2.0 for each cell, where D_bulb is theAttorney Docket No.: 064752-501001 WO maximum inner diameter of the bulb measured on a plane through the cell center and d_c is the interconnect inner diameter.

[0036] In some embodiments, emitter indexing to the cell pitch p satisfies a midplane registration error < 0.1 p for the positions of emitter centers relative to corresponding cell midplanes.

[0037] In some embodiments, for the six-emitter layout, the axial cube-LED centers are located on planes separated by the pitch p and are concentric with the cell center within Al < 0.1 R eff, and face normals are orthogonal to S3 within ±3°.

[0038] In some embodiments, the apparatus includes a diffuser optic on the emitter side selected from micro-structured, holographic, molded, or ground-glass diffusers having a scattering half-angle sufficient to reduce the coefficient of variation over {<!> > O*} to < 0.2 while maintaining 8 within the maximum normal-offset tolerance.

[0039] In some embodiments, for the two-emitter layout and the four-emitter layout, planar LED plates or LED strip columns further include reflective baffles or liners positioned to return spill light toward the inner wall.

[0040] In some embodiments, emitters remain outside the wetted volume behind windows bonded to the cell, the windows having anti-reflection coatings and the face-to-wall spacing set to 8 so that the iso-flux boundary remains coincident with the inner wall within the maximum normal-offset tolerance.

[0041] One embodiment of an assembly for photochemical reactions is a wireless cube LED assembly for illuminating photochemical flow cells. The assembly can include a polyhedral body having six faces arranged as three mutually orthogonal opposed pairs, six LED emitters, one on each face of the six faces, a receiver coil disposed within the body, a rectifier coupled to the receiver coil, a constant-current driver coupled to the six LED emitters, and alignment features configured to set a nominal optical standoff (8) and orthogonal orientation relative to a mating surface. The assembly can be configured to receive power inductively from a transmitter coil at a transmitter frequency (f_T).

[0042] In some embodiments, an edge length of the body is 5-20 mm. In some embodiments, an emission band of the LED emitters is 200-700 nm with a per-face radiant flux of 10-1000Attorney Docket No.: 064752-501001 WO mW at 25 °C. In some embodiments, an inductance of the receiver coil is 0.1-10 mH and the transmitter frequency is 100 Hz-100 kHz.

[0043] In some embodiments, the faces are fabricated on FR-4 or aluminum-core PCB and the body is filled or coated with potting selected from silicone or fluorosilicone and optionally parylene.

[0044] In some embodiments, the transmitter coil is coupled to a controller configured to sweep the transmitter frequency across a band with assemblies installed in a reactor and to select an operating frequency (f_T*) that maximizes received power or minimizes primary current for a specified secondary load.

[0045] In some embodiments, the assembly is configured to be inserted into keyed pockets of a flow-cell chain at planes separated by a pitch (p), and when installed the emitter face centers lie on an emitter-reference surface Sb defined as the outer normal offset of the inner wall by 5, with face normals orthogonal to Sb within ±1°.BRIEF DESCRIPTION OF THE DRAWINGS

[0046] These and other features will be more readily understood from the following detailed description taken in conjunction with the accompanying drawings, in which:

[0047] FIG. 1 A shows a cross-section of an isoflux cell for a 2-LED emitter setup;

[0048] FIG. IB shows a three-dimensional (3D) model of an isoflux cell for a 2-LED emitter setup;

[0049] FIG. 2A shows a cross-section of an isoflux cell for a 4-LED emitter setup;

[0050] FIG. 2B shows a 3D model of an isoflux cell for a 4-LED emitter setup;

[0051] FIG. 3 A shows a cross-section of an isoflux cell for a 6-LED emitter setup;

[0052] FIG. 3B shows a 3D model of an isoflux cell for a 6-LED emitter setup;

[0053] FIG. 4A shows a cross-section of a photoreactor with chains of 2-LED isoflux flow cells;

[0054] FIG. 4B shows a chain of three 2-LED isoflux flow cells;Attorney Docket No.: 064752-501001 WO

[0055] FIG. 5A shows a cross-section of a photoreactor with chains of 4-LED isoflux flow cells;

[0056] FIG. 5B shows a 4-LED isoflux flow cell with LEDs and their supports;

[0057] FIG. 6A shows a cross-section of a photoreactor with chains of 6-LED isoflux flow cells;

[0058] FIG. 6B shows 6-LED iso-flux flow cell chains;

[0059] FIG. 7A shows a wireless LED cube;

[0060] FIG. 7B shows an exploded view of the wireless LED cube of FIG. 7A;

[0061] FIG. 7C shows a flow cell with cube pockets for accommodating a wireless LED cube;

[0062] FIG. 8A shows a photoreactor with an induction coil;

[0063] FIG. 8B shows a cross-section of a photoreactor with wireless LED cubes and 6-LED isoflux flow cells;

[0064] FIG. 8C shows wireless LED cubes and 6-LED isoflux flow cell chains;

[0065] FIG. 9 shows a cross-section of a 6-LED isoflux flow cell with coated internal surfaces;

[0066] FIG. 10 shows an example computational synthesis workflow can be used in some embodiments to convert high-level photochemical requirements into manufacturable isoflux-cell geometry and operating setpoints;

[0067] FIG. 11 shows a block diagram of an example system for powering a wireless LED assembly; and

[0068] FIG. 12 shows a block diagram of an example computer system.DETAILED DESCRIPTION

[0069] There is a need for a geometrically defined, isoflux reactor architecture that: (i) produces near-uniform fields at the minimum per-source drive, (ii) scales by serial / parallelAttorney Docket No.: 064752-501001 WO composition under residence-time and AP constraints, and (iii) admits manufacturable implementations.

[0070] Described are photochemical reactors that deliver uniform photon flux at scale by composing three-dimensional arrays of isoflux flow cells into series cell chains and arranging multiple chains in parallel under controlled residence-time and pressure-drop constraints. Each flow cell can be defined by an illuminated internal surface coincident with a O = $* isoflux boundary and can be illuminated by opposed, orthogonally oriented LED emitters in 2-LED, 4- LED, or 6-LED topologies. The 6-LED topology can be realized either with compact wired cube LED modules or with wireless, inductively powered cube LED modules that eliminate in-process harnessing. Flow cells can be linked by short, small-diameter interconnects sized to minimize dark volume while meeting a chain pressure-drop limit. Parallel flow cell chains can be fed and collected by manifolds configured to enforce a uniform pressure gradient for even flow splitting. Embodiments include transparent, monolithic cells (e.g., SLA-printed), windowed polymer / metal / glass / ceramic constructions, and pocketed chains for press-fit cube LEDs.Optional surface functionalization of the isoflux walls (protective barrier, photoreagent, adsorbent) extends performance and durability. Some embodiments integrate wireless cube LEDs with an agitator rotor so that rotation provides both mixing and volumetric illumination. The architecture can scales by pitch-defined chaining and bundle tiling to satisfy a broad range of UV-visible photochemical applications.

[0071] Certain embodiments will now be described to provide an overall understanding of the principles of the structure, function, manufacture, and use of the devices, systems, and methods disclosed herein. One or more examples of these embodiments are illustrated in the accompanying drawings. Those skilled in the art will understand that the devices, systems, and methods specifically described herein and illustrated in the accompanying drawings are nonlimiting embodiments. The features illustrated or described in connection with one exemplary embodiment can be combined with the features of other embodiments. Such modifications and variations are intended to be included within the scope of the present subject matter.

[0072] Further, in the present disclosure, like-named components of the embodiments generally have similar features, and thus within a particular embodiment each feature of each like-named component is not necessarily fully elaborated upon. Additionally, to the extent that linear or circular dimensions are used in the description of the disclosed systems, devices, and methods, such dimensions are not intended to limit the types of shapes that can be used inAttorney Docket No.: 064752-501001 WO conjunction with such systems, devices, and methods. A person skilled in the art will recognize that an equivalent to such linear and circular dimensions can easily be determined for any geometric shape.Definitions and Symbols

[0073] The following definitions and symbols are examples and are not intended to limit the present disclosure.

[0074] O* : target photon-flux threshold.

[0075] a: attenuation (absorption) coefficient of the process medium at the operating wavelength.

[0076] Iso-flux boundary: the locus of points where photon-flux density equals the target photon-flux threshold O* under the stated illumination layout and optical standoff 8.

[0077] Photon-flux density model: an attenuation model that accounts for absorption (a) at the operating wavelength and emitter angular emission.

[0078] 0 (theta): off-axis emission angle measured at the emitter, between the emitter’s inward optical axis and the vector to the evaluation point

[0079] 8: optical standoff distance from an emitter face to the outer reference surface used in the optical model.

[0080] S8 (emitter-reference surface): the outer normal offset of the inner wall by distance 8; nominal placement surface for emitter face centers / normals.

[0081] R_eff: effective illuminated radius, R_eff = (3 • V_illum / 47i)A(l / 3), where V_illum is the illuminated volume enclosed by the iso-flux boundary.

[0082] r min: minimum local radius of curvature of the inner wall.

[0083] Dmin(minimum internal dimension): manufacturing-set lower bound on cavity size, defined as the narrowest clear chord between opposing points of the interior boundary (measured along the symmetry locus through the minimum-flux region); Dmin fixes the cavity scale from which the center-to-emitter distance and interior contour are derived.Attorney Docket No.: 064752-501001 WO

[0084] d_c, L_c: interconnect capillary inner diameter and length, respectively.

[0085] p: axial repetition interval (pitch) of cells / cube-LED planes along a chain.

[0086] s: lateral pitch between neighboring chains / LED carriers in plane view.

[0087] CV over {<D > <D* } : coefficient of variation of photon-flux density computed on the voxel set meeting or exceeding 0*.

[0088] Optical plane (for windows / encapsulants): the emitter-facing planar surface of a window or transparent encapsulant intended to be co-planar with the inner wall within the stated tolerance.

[0089] Switchable dipolar surface layer: a surface coating whose effective dipole / affinity reversibly changes (e.g., spiropyran / merocyanine or azobenzene systems) to modulate adsorption for chromatography or capture / release operations.

[0090] LED array centerline / chain axis: best-fit line through emitter centers on a plate / strip / column (array centerline) and through cell centers along a chain (chain axis).Isoflux flow cell geometry

[0091] FIGS. 1A-1B, 2A-2B, and 3A-3C show example isoflux flow cells for 2-LED, 4- LED, and 6-LED emitter setups, respectively.

[0092] As shown in FIGS. 1A-1B, the flow cell can be a light-transmissive vessel whose interior boundary (FIG. 1A, 100) is shaped to substantially follow an isoflux surface generated by a prescribed, symmetric arrangement of solid-state light emitters (FIG. 1A, 103) mounted outside the vessel at a uniform standoff from the wall (FIG. 1 A, 100) and aimed inward toward a common cell center C (FIG. 1A, 110).

[0093] The flow cell geometry can be defined by four device features: (i) a minimum internal dimension Dmin that fixes the narrowest chord of the cavity (FIG. 1A, 104); (ii) a predetermined location of the minimum-flux region (MFR) set by the emitter pattern’s symmetry (on the perpendicular bisector of a two-emitter axis (FIG. 1 A, 102); on the orthogonal bisectors for the four emitter setup (FIG. 2A, 200); and along diagonal loci (45° to the nearest axes) for the six emitter setup (FIG. 3 A, 300)); (iii) equalized emitter distances from C that produce a smooth, convex inner boundary enclosing the MFR (FIG. 1A, 101); and (iv) one or more windowAttorney Docket No.: 064752-501001 WO portions (FIGS. 1A-1B, 106) of the wall that are locally planar, flat, and oriented to face a corresponding emitter, the window being substantially normal to the emitter axis to aid efficient coupling of light into the cell, while the remainder of the interior boundary remains substantially coincident with the= <D* isoflux surface.

[0094] Emitter output (e.g., drive current and / or duty cycle) can be adjusted so that the local flux at the MFR (FIG. 1A, 102) meets a specified target minimum ( *), thereby fixing the isoflux boundary so that the interior wall substantially follows. Accordingly, the cavity maintains a substantially uniform radiant flux: low-flux regions are bounded by the specified target minimum 0*, and high-flux regions are curtailed by selecting emitter output to reduce peak intensity, thereby minimizing over-illumination inside the cell.Selection ofDmmfor isoflux flow cell geometry

[0095] In general, choosing a smaller Dmin can improve cell performance by shortening inmedium optical paths, reducing Beer-Lambert attenuation, lowering the emitter output required to meet the target minimum flux <b*, and flattening spatial gradients so fewer regions are overilluminated. However, in many cases, Dmin cannot be made arbitrarily small. Dmin is frequently bounded by packaging and fabrication limits, including (i) LED array pitch on the wiring package (typically between approximately 0.5-2.0 cm center-to-center), (ii) the optical standoff 5 (FIG. 1A, 105) needed for a flat window and coupling gap (typically approximately 0.1-0.5 cm; larger offsets can become bulky and inefficient), (iii) wall / window thickness and sealing strength, and (iv) manufacturing tolerances on the isoflux wall. Within these constraints, the design rule can be to select the smallest Dmin consistent with the chosen array pitch, 5, and build method. Representative embodiments use Dmin in the 0.5-5 cm range, which balances uniformity and power efficiency against practical packaging and manufacturability.Manufacturing variation of isoflux flow cell geometry

[0096] The as-built interior wall may deviate from the ideal O= >* isoflux surface due to practical process effects (print stair-stepping, shrink / warp, machining radius). To specify acceptable variation independent of topology, we define an effective cell radius RF= (37 / (4TT / 3))1 / 3, where V is the total internal cell volume; dimensional departures are then expressed relative to RF. In some embodiments, all geometric deviations from the ideal wall - including local radial offsets of the isoflux surface, window flatness error, and curvature departures - are limited to not more than 20% of the corresponding effective value, i.e.,Attorney Docket No.: 064752-501001 WOI Ar |<0.2 RF (and analogous bounds for linear / areal features scaled by RF). These bounds preserve the threshold guarantee by ensuring that any local perturbation of path length and view factor does not materially reduce the minimum photon flux below 0*. If a measured build approaches the limit, the emitter output may be trimmed upward to re-establish d>(MFR)><!>* while maintaining minimized over-illumination elsewhere.Modeling of isoflux flow cell geometry and performance

[0097] Governing irradiance model: Local irradiance at a given point (e.g., FIG. 1 A, 111) inside of an isoflux flow cell can be modeled by near-field superposition with angular falloff and Beer-Lambert attenuation (Equation (1)). (cos 0_i)we~a d-1(1)In Equation (1), 0_i is the off-axis emission angle (FIG. 1A, 112) measured at the emitter between the emitter’s inward optical axis and the vector to the evaluation point, T is the window transmittance / coupling factor (T G (0,1] is a unitless efficiency factor that accounts for power lost at the flat window portion and adjacent interfaces), an d_i is the distance from the emitter to a given point i (FIG. 1A, 113).

[0098] Uniformity Metric - Coefficient of Variation: Within the designed working domain fl = {x : <£»(%) > <P *}, uniformity can be quantified by the dimensionless coefficient of variation (CV), defined as the ratio (Equation (4)) of the spatial standard deviation of flux (a, Equation (3)) to its spatial mean (p, Equation (2)). / z = (l / y) f J2 <p(x) dVo2= (1 / 7) f jl (<P(x) - pf dVCV = a / p

[0099] In discrete evaluations, (simulation voxels or probe measurements), the spatial mean and spatial standard deviation of flux can be defined by Equations (5) and (6), respectively: p = (l / lV) ^(5)Attorney Docket No.: 064752-501001 WO<7 = [ (1 / W) £, ( 1 - g)2] (6)[000100] Because CV is scale-invariant, uniformly increasing or decreasing emitter output does not change CV. Thus, CV isolates geometric and angular uniformity from absolute intensity. Lower CV indicates a tighter dose distribution - improving conversion, selectivity, and batch-to-batch reproducibility - while also signaling reduced over-illumination (less energy wasted in hot spots).[000101] If the effective attenuation of photon flux changes by Aa, the solved drive can be updated to preserve <D=<D* on the shell boundary (Equation (7)).(7)[000102] Orthogonal sources and closed packing: Three mutually orthogonal axes exist in 3D. opposed pairs along these axes yield six inward normals (±x, ±y, ±z). In some embodiments, six can be maximum count of pairwise orthogonal LED directions. Using more than six directions may force non-orthogonal normals, increase mutual shading, and break closed packing of congruent shells.[000103] FIG. 10 shows an example computational synthesis workflow 1000 can be used in some embodiments to convert high-level photochemical requirements into manufacturable isoflux-cell geometry and operating setpoints. Given a target minimum photon-flux threshold O* and a selected emitter topology, the workflow 1000 computes the interior wall that substantially coincides with the <D = O* boundary and the minimum emitter drive necessary to achieve O* at the topology-defined minimum-flux region (MFR). The workflow 1000 is not intended to be limiting; in some embodiments, other numerical methods may be substituted to produce the same results.[000104] The workflow inputs can be design parameters supplied to the workflow 1000, including: the target threshold ©*, optical attenuation coefficient a of the reaction mixture, selected emitter topology (2-LED, 4-LED, or 6-LED), optical standoff 5, and Dmin per the geometry section.Attorney Docket No.: 064752-501001 WO[000105] The workflow outputs can be the design artifacts produced by the workflow 1000, including, for example: (i) the solved per-emitter drive 10* that enforces d>(P_MFR) = d>*; (ii) the O = <D* isosurface defining the interior wall; (iii) validation metrics (e.g., volume CV within {<D > (D*}, topology-appropriate slice CVs, <Dmin, ©mean, ©max); and (iv) manufacturable CAD files (e.g., STL / STEP with declared units and tolerances), along with 2D contour plots for drawings and emitter coordinates / aims.[000106] The workflow 1000 can be performed, all or in part, by one or more processors of a computer system. For example, the workflow 1000 can implemented as instructions stored in non-transitory memory of the computer system. Alternatively, or in addition, the workflow 1000 can be included in non- transitory computer readable memory storing the workflow 1000 as instructions which, when executed by one or more processors forming part of a computer system, causes the processor(s) to perform operations of the workflow 1000. Additional details regarding a computer system that can perform operations of the workflow 1000 are provided herein with respect to FIG. 12.[000107] The workflow 1000 is intended only as an example. In some implementations, a method of converting high-level photochemical requirements into manufacturable isoflux cell geometry and operating setpoints can include operations performed in a different order than the operations of the workflow 1000. In some implementations, a method of converting high-level photochemical requirements into manufacturable isoflux cell geometry and operating setpoints can include operations in addition to those of the workflow 1000. In some implementations, a method of converting high-level photochemical requirements into manufacturable isoflux cell geometry and operating setpoints can omit aspects of the workflow 1000.[000108] In some embodiments, the workflow 1000 includes, at 1005, characterizing the reaction. Characterizing the reaction can include measuring or estimating the mixture’s attenuation coefficient a at the operating wavelength(s). From kinetics and dose targets, the minimum photon-flux threshold * can be set, accounting for quantum yield, catalyst / photosensitizer extinction, and desired conversion / selectivity.[000109] In some embodiments, the workflow 1000 includes, at 1010, selecting topology and packaging constraints. 2-, 4-, or 6-LED symmetry can be chosen and optical standoff 8 and window location(s) can be defined in view of array pitch, module clearances, and thermalAttorney Docket No.: 064752-501001 WO constraints. Geometry selection (e.g., Dmin) can be handled in its dedicated section; use that value here or specify an equivalent radius r = Dmin / 2 + 3.[000110] In some embodiments, the workflow 1000 includes, at 1015, placing emitters.Opposed emitters can be located at ±r on the selected axes (±x, ±y; and ±z for 6-LED) and aimed toward the cell center. Each emitter can be associated with a flat window patch nominally normal to its optical axis for efficient coupling.[000111] In some embodiments, the workflow 1000 includes, at 1020, using point Lambertian sources, inverse-square falloff, and Beer-Lambert attenuation applied only to the in-medium segment of each source-to-point path. Air and wall segments can be treated as lossless. The total flux O(x) can be the sum of contributions from all emitters.[000112] In some embodiments, the workflow 1000 includes, at 1025, locating the topology-specific MFR. For 2-LED, the MFR may lie on the perpendicular bisector of the emitter axis. For 4-LED, the MFR may lie on an orthogonal bisector off the LED plane. For 6-LED, the MFR may lie along a body-diagonal locus (45° to each axis).[000113] In some embodiments, the workflow 1000 includes, at 1030, solving for a minimum emitter output. The smallest uniform per-emitter drive 10* can be determined such that O(P_MFR) = 0* at the MFR. If multiple wavelengths or channels are used, the minimum emitter output can be solved for jointly or sequentially to meet <b* while minimizing over-illumination elsewhere.[000114] In some embodiments, the workflow 1000 includes, at 1035, evaluate the flux field. Over a 3D grid that bounds the anticipated cavity, <b(x) can be evaluated using the radiometric model and 10*. A resolution sufficient to resolve the O = * surface (e.g., grid spacing < l%-2% of the effective radius) can be used.[000115] In some embodiments, the workflow 1000 includes, at 1040, extracting the isoflux wall. Compute the isosurface S = {x : (x) = O*}. Where specified, local segments can be replaced with flat window portions that are substantially normal to the corresponding emitter axis. S can be exported as a watertight surface.[000116] In some embodiments, the workflow 1000 includes, at 1045, validating uniformity and threshold. Validating can include confirming Omin > O* within the enclosed volume,Attorney Docket No.: 064752-501001 WO computing volume CV over { > *}, and reporting topology-appropriate slice CVs (XY / XZ / YZ or body-diagonal slices). A hot-spot magnitude / extent can be inspected to verify reduced over-illumination with 10*.[000117] In some embodiments, the workflow 1000 includes, at 1050, applying manufacturing adjustments. Wall thickness, fillets / draft, interconnects, and datum features can be added.Geometric departures from the ideal isoflux wall can be maintained within specified limits (e.g., not more than 20% of the effective radius RF where RF = (3 V / 4TC)A( 1 / 3)). Omin and CV can be rechecked after adjustments.[000118] In some embodiments, the workflow 1000 includes, at 1055, generating or provided deliverables. In some embodiments, providing deliverables can involve providing STL / STEP (declared units), emitter coordinates / aim vectors, 2D flux contour plots for drawings, and / or a tabular report of min / mean / Omax and CV (volume and key slices). The solved 10* and recommended operating drive (current / duty) can also be included.[000119] In some embodiments, the workflow 1000 includes, at 1060, performing operational trimming. A procedure to trim emitter output can be provided post-build to re-establish <D(P_MFR) > O* if as-built deviations are detected, while keeping over-illumination minimized.[000120] Equivalent solvers (ray-based, FEA / radiative transfer, or GPU path-integration), different die sizes, or wavelength bands may be used without departing from the workflow can all produce an interior wall substantially coincident with <D = <D* for the given inputs.Topology Selection Guidance[000121] In some embodiments, the 2-LED geometry may include fewer emitter s / dri vers and may be easily sealed, relative to the 4-LED and 6-LED topologies. In some embodiments, the 2- LED geometry can be used when the reaction tolerates modest non-uniformity, when simplicity and low parts count outweigh incremental energy savings, and when higher per-LED current and associated thermal management are acceptable.[000122] In some embodiments, the 4-LED topology can improves uniformity with moderate per-LED power and straightforward packaging (orthogonal strips or plates). The 4-LED topology can enable reduced per-LED currents and better wall-plug efficiency without the complexity ofAttomey Docket No.: 064752-501001 WO the 6-LED topology, and therefore can be used when manufacturability and uniformity are co- priorities.[000123] In some embodiments, the 6-LED geometry offers the lowest CV and enables the lowest per-LED drive, maximizing electrical / optical efficiency and minimizing thermal load - especially at higher attenuation a. The 6-LED topology can be used when uniformity is paramount, energy efficiency is critical (heat-sensitive substrates, expensive catalysts), or when scaling to dense 3-D arrays with wireless LED cubes and coaxial solenoidal coupling to preserve tight packing and eliminate wiring shadows.Numerical examples and comparison to plate and tubular reactors[000124] This section compares representative isoflux flow cell topologies (2-LED, 4-LED, 6- LED) against conventional plate and tube reactors using a common optical model and target, showing that the 4-LED and 6-LED cells achieve the specified minimum flux with lower emitter drive and greater spatial uniformity. The numerical values used in this comparison were selected to reflect a typical visible-light photoredox reaction under blue irradiation (e.g., approximately 450-1270 run). For such systems, a working formulation at practical concentrations exhibits a moderate attenuation on the order of a ~ 1 cm-1; campaigns targeting high single-pass conversion commonly set a minimum photon-flux threshold of C>* = 20 mW cm-2at the topology’s minimum-flux region to guarantee the required dose. The standoff of 8 = 0.2 cm reflects a compact, manufacturable window thickness plus coupling gap for blue LED packages. Dmin of 1 cm internal span (for isoflux cells) or 1 cm gap / diameter (for plate / tube comparators) represents a module suitable for small scale manufacturing. These choices are illustrative and non-limiting; other wavelengths, mixtures, and targets can be evaluated identically within the same workflow.[000125] Beer-Lambert attenuation may be applied only inside the medium. Air and wall segments can be treated as lossless. Emitters can be modeled as finite-area Lambertian dies (1 x1 mm) placed at a uniform optical standoff 8 from the exterior wall / window and aimed inward.[000126] Parameters: attenuation a=l cmA-l; threshold <I)*=20 mW / cm'2; standoff 8=0.2 cm; Dmin=l cm for isoflux cells; plate gap 1 cm; tube diameter 1 cm; plate / tube LED strip pitch 1.5 cm.[000127] For each topology, the per-emitter radiance L0 (mW sr-1cm-2) can be solved so that the topology-specific MFR attains <I)(PMFR)=*. For the plate and the tube, L0 can then min-Attorney Docket No.: 064752-501001 WO bounded so the global minimum in the evaluation window satisfies 0min>d>*. Uniformity can be summarized by the slice CV on a topology-appropriate slice (diagnostic 2D section through the domain), which upper-bounds the full- volume uniformity trend. The finite die area yields an equivalent point-source intensity 10 =L0x0.01 cm2for easy comparison.[000128] Results of the comparison are provided in Table 1.[000129] Interpretation of results: Relative to conventional reactors, the 4-LED and especially the 6-LED isoflux cells reach the same threshold O* with substantially lower emitter drive (e.g., 6-LED LO-364 vs. -1288 and 1693 for tube / plate after min-bounding) and exhibit markedly lower CV (greater flux uniformity). The 2-LED cell improves over plate but is surpassed by the 4- and 6-LED configurations. Lower required drive reduces energy consumption and thermal load; lower CV narrows the dose distribution, improving conversion / selectivity and reproducibility.[000130] Practical significance: Because the isoflux wall is the <D=<D*boundary and emitter output is set to the lowest level that meets the threshold at the MFR, the design eliminates darkAttorney Docket No.: 064752-501001 WO pockets while curtailing peaks, delivering uniform, specification-bounded illumination with less over-illumination than plate / tube geometries. For concentrated (higher-a) formulations, the advantage of the 4- and 6-LED cells increases: added orthogonal emitters shorten effective paths to under-served regions and suppress hot-spot dominance near any single emitter.[000131] The tabulated CVs are 2D diagnostic values on topology-appropriate slices; fullvolume CVs are typically lower (best case reduction with 6-LED) but follow the same ranking. Equivalent results are obtained with finite-area or point-source formulations that apply attenuation only in the in-medium segment.Cell chain design, manufacturing, and mounting[000132] Cell chains (FIGS. 4A-4B, 400) can be formed by connecting illuminated isoflux cells (401) with short cylindrical interconnects (402). In some embodiments, because the interconnects are weakly illuminated, their internal volume is minimized while still meeting hydraulic and manufacturing requirements. Interconnect length L_c can be selected to register the emitter planes with the centers of the adjacent cells at the chosen pitch p, and interconnect inner diameter d_c can be selected to control pressure drop, reduce clogging risk, and satisfy manufacturing tolerances. In some embodiments, d_c does not exceed 2 mm and L_c does not exceed 30 mm. In some embodiments, the chain design can be subject to two performance constraints and one manufacturing constraint. The performance constraints can be a target residence time and a maximum allowable pressure drop across the chain. The manufacturing constraint can be a maximum manufacturable chain length L_max for the selected process. For example, when an industrial stereolithography 3D printer with an effective build length of about 100 cm is used, the chain length L chain can be set to be less than or equal to 100 cm.[000133] In some embodiments, pitch p is the axial length per cell and includes the illuminated axial thickness t_bulb, the interconnect length L_c: p = t_bulb + L_c. With t_bulb = 1 cm and L_c = 0.5 cm, gives p = 15 mm. In some embodiments, the series count satisfies N less than or equal to floor of L max divided by p. In some embodiments, with L max = 100 cm and p = 1.5 cm, N is at most 66. In some embodiments, the chain length is L_chain = N t_bulb plus (N minus 1) L_c. For N = 100, L_chain = 99 cm, which satisfies the SLA length cap.[000134] In some embodiments, V cell is the illuminated hold up volume per cell and V ic is the dark interconnect hold up. For an interconnect of inner diameter d_c, the cross sectional area can be A_c = n d_c2divided by 4 and the dark hold up can be V_ic = A_c L_c. Total hold up canAttorney Docket No.: 064752-501001 WO be V_chain = N V_cell + (N minus 1) V_ic. The residence time constraint can be r_chain = V_chain divided by Q greater than or equal to r_target. The residence time bound gives Q_res(N) = V_chain divided by r_target.[000135] In some embodiments, for flow Q through one interconnect, the pressure drop can be AP_ic = [ f_c times (L_c divided by d_c) plus K_ic ] times p v_c2divided by 2, v_c = Q divided by A_c, Re_c = p v_c d_c divided by p, f_c = 64 divided by Re_c for laminar segments and f_c approximately 0.3164 times Re c to the power of negative 0.25 otherwise. K_ic can be a lumped junction loss for one expansion into the bulb and one contraction back into the interconnect. In some embodiments, K_ic can be selected within 0.8 to 2.5 based on edge radius and taper. In some embodiments, a minimum edge radius of 0.15 mm can be specified to maintain K_ic less than or equal to 1.5. In some embodiments, a chain pressure drop is AP chain = (N minus 1) AP ic, and satisfies AP chain less than or equal to AP limit.[000136] In some embodiments, the illuminated body can be printed as a single transparent resin shell, so no discrete windows are used. LED arrays can be mounted outside the process volume and emit through the transparent wall at a fixed standoff 8. The illuminated hold up per cell can be V_cell = 0.32 mL for the 4 LED isoflux geometry. In some embodiments, with t_bulb = 7 mm, the effective illuminated cross section is A_bulb = V_cell divided by t_bulb = 0.457 cm2and the equivalent radius is R_eff = 3.82 mm. In some embodiments, the interconnect satisfies d_c less than or equal to 0.2 R_eff, which is about 0.76 mm, unless A_bulb is increased by optical redesign. In some embodiments, the resin is selected for high transmittance at the LED band, low yellowing under exposure, and chemical compatibility with the process fluid.[000137] In some embodiments, interconnect diameter d_c is held to plus or minus 0.05 mm and interconnect length L_c is held to plus or minus 0.2 mm. In some embodiments, SLA shrinkage is compensated in the build file so that as printed d_c and L_c meet these tolerances after post cure. In some embodiments, each chain is printed as a single part that contains the repeating illuminated bulbs and the 3 mm interconnect necks at a pitch of 10 mm over a total length less than or equal to 100 cm. Post processing can include solvent rinse, UV post cure, thermal condition, and optical finishing of the inner surface by gentle honing or polishing to reduce scatter. If needed for solvent resistance or oxygen barrier, thin conformal coatings such as parylene or silica like barrier films can be applied to the inner surface.Attorney Docket No.: 064752-501001 WO[000138] In some embodiments, assuming: water at 25 °C with p = 997 kg m-3and , = 0.00089 Pa s, V_cell = 0.32 mL, t_bulb = 7 mm, L_c = 3 mm, N = 100, r_target = 5 minutes, AP_limit = 30 psi, K_ic = 1.5, the illuminated cross section gives R eff = 3.82 mm, so d_c can be selected at or below 0.76 mm to satisfy d_c less than or equal to 0.2 R eff. Results at Q = Q res can include:• Case A, d_c = 0.50 mm. A_c = 0.196 mm2. V_ic = 0.000589 mL. V_chain = 32.058 mL. Q_res = 0.006412 L min-1. Re_c = 305. f_c = 0.210. AP_chain = 5.851 psi.• Case B, d_c = 0.60 mm. A c = 0.283 mm2. V ic = 0.000848 mL. V chain = 32.084 mL. Q res = 0.006417 L min-1. Re c = 254. f_c = 0.252. AP chain = 2.825 psi.• Case C, d_c = 0.70 mm. A_c = 0.385 mm2. V_ic = 0.001155 mL. V_chain = 32.114 mL. Q res = 0.006423 L min-1. Re c = 218. f_c = 0.294. AP chain = 1.527 psi.[000139] The aforementioned cases A-C meet the 5 minute residence time and remain well below the 30 psi limit. Flow in the interconnects can be laminar in each case.[000140] Residence time can be verified by a nonreactive tracer step test using conductivity or UV tracer, with hold up computed from the first temporal moment. Pressure drop can be verified by water flow at 25 °C while recording differential pressure at Q_res. Optical compliance can be verified by measuring through wall transmittance at the LED wavelength before and after an exposure test. In some embodiments, acceptance requires r_chain greater than or equal to r target, AP chain less than or equal to AP limit, and transmittance above a specified threshold.[000141] The design method can be applied for fluids with viscosity between 0.3 and 10 mPa s and density between 800 and 1300 kg m-3, with f_c and Re_c evaluated at process temperature. If a larger d_c is desired for hydraulic reasons, the illuminated cross section can be increased so that d_c remains at or below 0.2 R eff. In some embodiments, if L max must be exceeded, the chain can be split into two printed segments joined by a straight sleeve of length 10 to 15 mm and inner diameter equal to d_c within plus or minus 0.02 mm so that added minor loss per splice is less than 5 percent of one interconnect.Attorney Docket No.: 064752-501001 WO[000142] In some embodiments, the cell chain is implemented with different combinations of materials and manufacturing methods selected to match the application. Selection criteria include solvent set, UV or visible band, required residence time, allowable pressure drop, pressure rating, temperature, fouling tendency, cleanability, and regulatory needs. The optical and hydraulic design method may be unchanged, but the materials of construction and the way the geometry is produced differ to meet the requirements.[000143] In some embodiments, the process wetted body uses PVDF, PFA, PTFE, FEP, PPSU, PEEK, COC, borosilicate glass, fused silica, or alumina class ceramics. In an additional embodiment, structural jackets and LED carriers use aluminum alloys, anodized aluminum, copper, aluminum nitride, stainless steels, or ceramics. Wall thickness can be chosen to satisfy the pressure rating with a factor of safety and to meet manufacturability constraints such as minimum draw thickness and dimensional stability.[000144] In some embodiments, the fluid body is injection molded in PVDF, PPSU, PEEK, COC, or FEP class materials or machined using CNC methods as two halves with interconnect seats and alignment features. In some embodiments, transparent windows of fused silica, borosilicate, or COC are bonded into pockets to increase optical efficiency when the base polymer has higher absorption at the LED wavelength. Windows can be anti reflection coated on the air side to reduce Fresnel loss and are sealed with compatible adhesives or welded frames.[000145] In an additional embodiment, the fluid body is machined by CNC or other precision machining methods either as two mating halves with interconnect seats and alignment features or as a monolithic block that is cavity milled and then finish reamed. Representative operations include three axis and five axis milling, boring, microdrilling, reaming, and internal honing to set the interconnect diameter and surface quality; optional diamond finishing is used on transparent surfaces to reduce scatter. In a further embodiment, transparent windows of fused silica, borosilicate, or COC are bonded into pockets to increase optical efficiency when the base polymer has higher absorption at the LED wavelength. Windows can be anti reflection coated on the air side to reduce Fresnel loss and are sealed with compatible adhesives or welded frames.[000146] In some embodiments, the process wetted body is glass or fused silica formed by flame work or diffusion bonding, or alumina class ceramic. LEDs can be outside the process volume and couple through fused silica windows that provide high transmittance and thermalAttorney Docket No.: 064752-501001 WO stability. This can be used for aggressive solvents, elevated temperature duty, or very low extractables.[000147] In some embodiments, a metal or metal lined shell provides structural strength, with fused silica or COC windows for optical coupling. Internal reflective finishes such as protected aluminum, protected silver, or multilayer dielectric mirrors are used on non window surfaces to improve field uniformity. The windows can be configured to allow efficient coupling while the metal body provides robustness and heat rejection.[000148] In some embodiments, non window internal surfaces receive protected metal mirrors, dielectric stacks tuned to the LED band, or diffuse PTFE class inserts. These finishes can increase optical efficiency by directing stray light back into the illuminated volume and by smoothing local gradients. Coatings can be masked at seal lands and protected with top coats compatible with the process fluid.[000149] In some embodiments, the manufacturing route can be selected by a decision method that takes as inputs the solvent set, the target residence time, the chain pressure limit, the required transparency, the expected temperature, and the allowable length envelope. The method can choose among monolithic transparent SLA, injection molding with or without windows, blow molding or extrusion with thermoformed bulbs, glass or ceramic bodies with windows, and metal lined bodies with windows. The method can then sets pitch, interconnect diameter, and wall thickness to meet hydraulic and optical constraints while staying under the manufacturing length cap.[000150] In some embodiments, windows are used to increase efficiency when the base body material absorbs at the LED wavelength or when long term exposure could cause yellowing. Windows with high transmittance and anti reflection coatings can reduce optical loss and allow a broader choice of structural materials. Windows can also allow independent replacement or cleaning without disturbing the structural body. In some embodiments, where the printed body is sufficiently transparent, windows can be omitted as in the monolithic transparent SLA embodiment.[000151] In some embodiments, every embodiment is validated by the same two constraints and the same test methods. Residence time can be verified by a tracer step test at the target flow and pressure drop can be verified by water flow at 25 °C at that flow. Optical compliance can be verified by measuring transmittance at the LED band through the body or through the windowsAttorney Docket No.: 064752-501001 WO as applicable. Acceptance requires T_chain greater than or equal to z_target, AP_chain less than or equal to AP_limit, and optical transmittance above the specified threshold.[000152] In some embodiments, the inlet header (FIG. 4, 408) and the outlet header (409) are each formed as a shallow dished bottom plenum in a single pass shell style. In some embodiments, the dish provides a rapid cross sectional expansion that creates a gentle pressure gradient while keeping the axial length short. In some embodiments, flow enters the dished inlet (409), expands radially, passes a perforated equalization headplate (410) that sets a small controlled pressure drop, and then enters the branch takeoffs to the cell chains (411). In some embodiments, takeoffs are arranged on one or more pitch circles. In some embodiments, the geometry maintains AP_chain greater than or equal to three times AP_header, range at the design flow while minimizing dead zones and end length.[000153] A reactor can be assembled in the manner of a single pass shell type heat exchanger. A headplate (410) can act as a tubesheet analog and carries the cell chains as a bundle. In some embodiments, each chain terminates in an inlet collar and an outlet collar that register to ports (411) in the headplate. In some embodiments, the dished inlet head (408) bolts to the headplate on one side and the dished outlet head (409) bolts to the opposite side to give precise alignment, short end chambers, and a predictable header gradient for uniform splitting among chains.[000154] In some embodiments, the headplate is sealed to the dished heads by gaskets or by adhesive bonds compatible with the materials of construction. Face seal grooves may accept O rings, encapsulated elastomer rings or glued gaskets. In some embodiments, when the cell chains are produced by 3D printing, the headplate is printed as an integral feature of the channels and the dished heads are printed or machined and bonded directly to the printed headplate to eliminate separate gaskets while preserving the designed pressure profile and the short overall length.LED array design and installation for 2-LED isoflux cell chain embodiments[000155] LED plate integration for 2-LED cell chains: In some embodiments, each cell chain (FIG. 4, 400) is sandwiched between two planar LED plates (403,404) held at a fixed optical standoff 8 from the transparent cell wall. The plates can be arranged back-to-back so that one plate (403) illuminates one side of a chain and the opposing plate (404) illuminates the adjacent chain (405) on the next row. Rows repeat to form a plate-chain-plate-chain stack. The LED plates can be 2D arrays on metal-core PCBs or AIN tiles fastened to liquid-cooled bases. In someAttorney Docket No.: 064752-501001 WO embodiments, standoffs and dowel features on the bases set 5 and keep parallelism within ±0.2 mm across the bundle.[000156] LED plate construction and cooling: In some embodiments, each LED plate comprises a constant-current LED matrix, a metal-core substrate, and a heat spreader bonded to a cold plate with coolant channels. Coolant lines (406) can be routed on the back of the plates and manifolded at the bundle perimeter. Junction-to-coolant thermal resistance can be maintained below a specified limit by choosing plate copper thickness, via density, and TIM bond-line thickness.[000157] Cell cooling corridors between rows: In some embodiments, the spacing between opposing LED plates that face a chain (407) defines a coolant corridor of height g_cool. A secondary liquid loop can be run through these corridors to remove process heat from the cells. Corridors are sealed from the process volume and from LED plate coolant. In some embodiments, flow guides or low-K baffles ensure uniform corridor velocity while preserving optical standoff. Condensation control can be provided when corridor coolant is below ambient dew point.LED array design and installation for 4-LED isoflux cell chain embodiments[000158] As shown in FIGS. 5A-5B, in some embodiments, each cell chain (501) is illuminated on four sides by four linear LED strips (502) positioned at a fixed standoff 8 from the transparent cell wall (503). Individual LEDs (504) on each strip can be placed to align with the cell positions set by the optical design so that every 4-LED isoflux cell receives the intended opposed and orthogonal pairing. The four strips around a given chain can be indexed to common datums so that 8 and angular aim repeat within tolerance across the bundle.[000159] In some embodiments, the four LED strips form a rectangular “four strip column” around a central spine (505). The spine can be a rod or a hollow tube that may be round, square, or rectangular. A rectangular tube can simplify flat mounting pads and preserve strip planarity. The spine can provide stiffness, a protected routing path for wiring, and optional internal coolant passages. Coolant can flow through the spine and can be manifolded at the bundle perimeter so that LED junction temperature stays within specification.[000160] In some embodiments, LED strips are mounted by adhesive bonding to prepared pads on the spine, by mechanical fasteners into molded or machined bosses, or by snap-in rails thatAttorney Docket No.: 064752-501001 WO capture a terminal base at defined locations. Fiducial features on the spine can set the z-height and yaw of each strip. In some embodiments, tolerances are ±0.2 mm for standoff 8 and ±1 degree for strip yaw. Shims or adjustable rails can be used where needed to fine time overlap at cell boundaries.[000161] In some embodiments, the architecture repeats in a checker lattice. Each cell chain can be surrounded by four LED strips on its four faces, and each four strip column is surrounded by four cell chains. In some embodiments, pattern tiles in x, y, and z at the chain pitch so that columns and chains alternate through the module. In some embodiments, headplate registers both columns and chains so that the optical alignment is preserved over the full length under the manufacturing length cap.[000162] In some embodiments, the gap between the LED strips and the neighboring cell faces (506) is used as a sealed coolant corridor to remove process heat from the cells. Corridor coolant can be hydraulically isolated from both the process fluid and the LED spine coolant. Baffles or flow guides can distribute corridor flow while keeping clear optical paths. Condensation control and leak detection can be provided when corridor coolant temperature is below ambient.LED array design and installation for 6-LED isoflux cell chain embodiments[000163] In some embodiments, the LED emitters Eire positioned at six orthogonal locations per cell. Two alternative embodiments are disclosed. A wired cube LED (FIGS. 6A-6B) can be used when the smallest cube envelope and the smallest 8 are required and the harness count remains manageable. A wireless cube LED (FIGS. 7A-7C) can be used for large scale arrays where process side harnessing and feedthroughs would create routing and service complexity.[000164] Wired cube: In some embodiments, the cube LED (FIGS. 6A-6B, 601) is a compact module with six emitters (602) on orthogonal faces and external wiring (603). Because no receiver coil or on board driver is included, the cube edge and wall thickness can be reduced, enabling a smaller 8 and tighter optical packing. In some embodiments, harness trays and keyed connectors manage wiring for the intended array size.[000165] Wireless cube: In some embodiments, a wireless “cube LED” (FIG. 7A, 701) integrates six emitters and is inductively powered. Each cube can have six face boards with LED emitters (FIG. 7B, 702), a centrally located receiver coil (FIG. 7B, 703), a rectifier, and a constant current driver assembled on an internal core, followed by potting or conformal coating.Attorney Docket No.: 064752-501001 WOThe cube form fixes orthogonality and 5 so all six faces illuminate the isoflux shell without any wiring inside the process volume.[000166] Inductive power and frequency tuning: Tuning can be performed by sweeping the transmitter coil (FIG. 8A, 801) frequency until the most efficient energy transfer is achieved with the receiver coils (FIG. 7B, 703) inside of wireless LED cubes. In some embodiments, a controller varies the carrier frequency f_T across an allowed band and measures a coupling metric such as received DC power at a cube, primary coil current at fixed transmitted power, or reflected power. In some embodiments, the optimal frequency f_T* maximizes received power or minimizes primary current for a specified secondary load. Because the effective inductance and capacitance depend on the number of receiver coils, their spacing, and their orientation, f_T* can shift with bundle geometry and with cube count. In some embodiments, tuning sweep is executed after assembly and stored per bundle; if geometry changes, the sweep can be repeated. Optional closed loop trim maintains operation near f_T* during temperature drift within electromagnetic compatibility limits.[000167] Flow cell chains prepared for cube integration: Flow cell chains can be manufactured by the same methods as the 2-LED and 4-LED chains (for example, transparent SLA print or molded halves) but with dedicated “cube pockets” (e.g., FIG. 7C, 704) at the design planes. Additionally, the flow cells can incorporate elongated channel (e.g., FIG. 7C, 705) arranged at an oblique angle so that they curve around the LED cube[000168] Odd and even chain variants for octahedral sequencing: In some embodiments, because the six face pattern repeats in alternating planes along the chain, two chain layouts are defined. A Type O (e.g., FIG. 6A-B, 604; FIG. 8B, 802) chain can have cube pockets on planes 1, 3, 5, etc., at the cell pitch. A Type E (e.g., FIGS. 6A-6B, 605; FIG. 8B, 803) chain can have pockets on planes 2, 4, 6, etc., offset by one half pitch. Alternating Type O and Type E chains in a bundle can interleave cube planes to maintain uniform optical spacing and avoid mutual shadowing.[000169] In some embodiments, pre-made cube LEDs are inserted into the prepared pockets and locked by snap features, solvent bonded tabs, or small fasteners into molded bosses. Placement may be by hand, pick and place, or jigs that reference the pocket keys. In some embodiments, for wireless cubes, no harness penetrates the process volume. A 8 gauge can verify standoff at selected planes during assembly.Attorney Docket No.: 064752-501001 WO[000170] Bundle assembly and external power delivery: Chains with integrated cubes can be mounted to the headplate and connect to the single pass dished inlet and outlet headers in the same manner as other cell chain bundles. One or more external transmitter coils can be installed around the chain or the bundle perimeter. In some embodiments, headplate features register coil position so coupling is uniform along the full chain length.[000171] Optional cooling embodiments: In some embodiments, liquid coolant flows in sealed corridors between cube faces and adjacent cell surfaces (FIG. 8, 804). The corridor loop can be hydraulically isolated from the process fluid and from any separate loop that cools external structures. In an additional embodiment for wireless cubes, the cube includes an internal heat spreader or miniature liquid passage. For wired cubes, drivers are located on the dry side and may be coupled to conventional heatsinks.[000172] Manufacturing sequence: In some embodiments, manufacturing can include (a) fabricating and testing cube LEDs by PCB and SMT, potting, and photometry; (b) manufacturing flow cell chains with Type O or Type E cube pockets; (c) inserting and locking cubes at the specified planes; (d) mounting chains to the headplate as a bundle; (e) for wireless embodiments, installing and tuning the transmitter coil or coils; and (f) verifying 8, coupling, residence time at Q_res, and AP_chain at the design limit.Wireless cube LED assembly for six-axis illumination[000173] In some embodiments, a wireless cube LED apparatus provides six-axis illumination from a compact polyhedral body. In some embodiments, the apparatus can include a cube or rectangular prism with edge length between 5 and 20 mm, six LED emitters fixed on mutually orthogonal faces, a centrally located receiver coil, a rectifier, and a constant-current driver. In some embodiments, alignment keys on the body set an optical standoff 8 to a target surface and hold face orthogonality within tolerance.[000174] In some embodiments, the body is a clear or translucent housing that encloses the electronics and incorporates molded or printed hard stops to define 8. In some embodiments, wall thickness is between 0.6 to 1.2 mm. Suitable materials include PETG, COC, or transparent SLA resins. In some embodiments, face substrates are FR-4 or aluminum-core PCBs bonded to internal pads. Optional studs or snap features can be provided for retention to a mating pocket. Sealing can be provided by conformal coating or potting such as silicone or fluorosilicone, with an optional parylene C barrier.Attorney Docket No.: 064752-501001 WO[000175] In some embodiments, each face carries one or more surface-mount LEDs. An example embodiment emits at 450 nm. In some embodiments, the supported spectral range spans 200 to 700 nm. In some embodiments, per-face radiant flux at 25 °C is selectable between 10 and 1000 mW. Optional diffusers or thin lenses on a face tailor the emission cone while maintaining total flux. In some embodiments, spectral mixing may be implemented by populating different faces with different wavelengths within 200 to 700 nm.[000176] In some embodiments, the receiver coil has inductance L between 0.1 and 10 mH and is coupled to a tuning capacitor to form a resonant input network. The coil may be wound with litz wire, fabricated as a printed spiral, or wound on a ferrite-assisted former. A rectifier and a current-mode driver can distribute regulated current to the six faces. The driver may include per- face balancing so that face currents remain within a specified tolerance.[000177] In some embodiments, the apparatus is configured to operate with a transmitter frequency f_T between 100 Hz and 100 kHz. The internal resonant network can be adjustable by component selection so that the receiver presents an efficient load within this band. Test pads can be provided for measuring received DC power or input impedance to support frequency selection. An optional temperature sensor on the core enables current derate functions implemented by the driver.[000178] In some embodiments, the apparatus includes an internal copper spreader bonded to the face boards to reduce thermal gradients. In some embodiments, miniature internal microfeatures or passages are formed within the body to enhance heat transfer to the housing. These features can be internal to the apparatus.[000179] In some embodiments, the body incorporates keyed edges, bosses, or grooves that mate with complementary features on a holder so that 8 is set within ± 0.2 mm and angular alignment is within ± 1 degree. A gauge surface on the apparatus can allow verification of 8 without disassembly.[000180] In some embodiments, the body is a rectangular prism with truncated comers while maintaining six emitting faces. The receiver coil may be single or dual. The apparatus may include electromagnetic shields, ferrite tiles, or spacing features to manage coupling in metal- proximate environments. The face optics may be clear, diffused, or lensed. The edge length is selectable within 5 to 20 mm to suit packing and optical standoff requirements.Attorney Docket No.: 064752-501001 WO[000181] FIG. 11 shows a block diagram of an example system 1100 for wirelessly powering an LED assembly (e.g., a cube LED assembly). In the illustrated embodiment, the system 1100 includes an LED assembly 1105, a transmitter coil 1130, and a controller 1140. The LED assembly includes LED emitters 110, a receiver coil 1115, a rectifier 1120, and a current driver 1125. The controller 1140 includes a processor 1145 and memory 1150.[000182] The controller 1140 can be implemented using any suitable device, for example a microcontroller, a laptop computer, a desktop computer, a server, or the like. The processor 1145 can include any suitable data processing device, for example a central processing unit (CPU), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or the like. The memory 1150 can include any suitable storage device. The memory 1150 can store instructions that, when executed by the processor 1145, cause the processor 1145 to control the transmitter coil 1130 to sweep a transmitter frequency to select an operating frequency that maximizes received power or minimizes primary current.Agitator-integrated wireless cube LED embodiment[000183] In some embodiments, wireless cube LED modules are integrated with a rotary agitator so that a single assembly provides both illumination and fluid mixing. The apparatus comprises a nonconductive shaft carrying paddles or hydrofoil blades. Wireless cube LEDs (edge 5 to 20 mm) are mounted on and within the blades at fixed orientations. Because power is transferred inductively, no wiring or slip rings are required on the rotating assembly.[000184] Transmitter coil geometry and axial alignment: In some embodiments, one or more stationary transmitter coils are arranged coaxially with the agitator shaft so that the magnetic field lines are aligned with the rotation axis. A solenoidal coil surrounding the vessel, or a stack of annular ring coils distributed along the liquid height, can produce an axially symmetric field. In some embodiments, with the cubes rigidly fixed on the rotor, axial alignment to the transmitter coil maintains strong coupling over rotation, so transmission efficiency remains high while the agitator turns.[000185] Wireless cube placement on mixing elements: In some embodiments, cubes are embedded flush into paddle faces or enclosed within streamlined fairings bonded to hydrofoils. Face normals can be set to a fixed triad relative to the blade chord and camber so that, as the rotor turns, each point in the fluid is swept by multiple orthogonal emitters. The pitch along theAttorney Docket No.: 064752-501001 WO shaft can be selected so adjacent blade levels do not shadow one another; a staggered pattern (odd / even axial planes) can reduce overlap and improves time-averaged uniformity.[000186] Mixing and illumination synergy: In some embodiments, the rotating assembly generates bulk circulation and local shear while the rotating light field averages spatial gradients. In some embodiments, although the instantaneous optical field is not an isoflux shell, the time- averaged irradiance approaches uniformity over the mixing time scale. This embodiment can be used where high Reynolds number agitation is required for mass transfer or suspension stability and where the process benefits from simultaneous strong mixing and volumetric illumination.[000187] Mechanical structure and retention: In some embodiments, the shaft and blades are fabricated from nonmagnetic, low-loss materials (e.g., COC, PVDF, PEEK, FEB, glass-fiber polymer composites) to minimize eddy losses and preserve coupling. Cube pockets can include keyed seats and hard stops that fix orientation; retention can be provided by snap lugs, adhesive bonds, or captive fasteners into molded bosses. All cube housings can be faired with fillets to avoid stagnation zones and to limit drag and vibration. The rotor can be dynamically balanced after cube installation.[000188] Power reception, regulation, and tuning in a rotating environment: In some embodiments, each cube includes a receiver coil (0.1 to 10 mH), rectifier, and constant-current driver as previously described. In some embodiments, a controller executes a frequency sweep of the stationary transmitter coil(s) over 100 Hz to 100 kHz with the rotor installed and at operating speed. The operating frequency f_T* can be selected to maximize received power or minimize primary current. In some embodiments, because the number and spacing of receiver coils on the rotor set the effective loading, f_T* is stored per agitator configuration; if blade count, cube count, or coil geometry changes, the sweep can be repeated.[000189] In some embodiments, the transmitter can be segmented into two or more axial zones that can be driven independently. In some embodiments, zoning allows local trim where vessel geometry or baffles alter coupling, and permits dimming profiles along the height to tailor dose. Zone transitions can be overlapped to avoid dark bands as the rotor sweeps past.[000190] Materials, optics, and sealing: In some embodiments, cube faces are clear or diffused according to the process wavelength (e.g., 200 to 700 nm). In some embodiments, the blade skins over embedded cubes are windowed or kept optically thin in front of each face to minimize loss. In some embodiments, wetted adhesives, coatings, and housings are selected for opticalAttorney Docket No.: 064752-501001 WO transmission at the target band and chemical compatibility. In some embodiments, where needed, thin protective coatings on the cube faces are applied as disclosed elsewhere in this application while preserving face flatness and orientation.[000191] In some embodiments, hollow shafts or internal blade channels carry a liquid coolant to manage temperature of the rotor assembly. In some embodiments, external coolant flows in narrow corridors between the fairings and the bulk fluid, guided by low-K baffles formed into the blade surfaces. Cooling features are optional and may not require wiring or slip rings.Surface functionalization of isoflux cells[000192] In some embodiments, coatings are to be applied to the surfaces of the isoflux cells, as shown in FIG. 9. The coating architecture can include: (a) a protective barrier that is chemically resistant and optically transmissive at the operating wavelength (901); (b) a photoreagent layer that harvests incident light to generate reactive intermediates (902); and (c) an optional adsorbent layer that concentrates target molecules near the photoreagent to shorten transport paths and improve kinetics (903). In some embodiments, where LEDs are inserted directly into an isoflux cell, the same coating families may be applied to the exposed LED surfaces (faces / windows) provided optical standoff 5 and alignment are maintained.[000193] In some embodiments, isoflux cell substrates include transparent SLA resins, COC / COGS, PVDF, PEEK, FEB, glass, or ceramic. Surfaces can be prepared by solvent cleaning and activation (e.g., corona or oxygen plasma) followed by adhesion promoters compatible with the chosen coating family. For low-energy polymers, a silica-forming primer (sol-gel) can provide surface -OH groups for subsequent silane coupling and for anchoring inorganic photoreagents.[000194] In some embodiments, protective layers comprise fluoropolymers, cyclic olefin copolymers (COC), polyacrylics / epoxies, PMMA, or polycarbonate selected for optical transmission and chemical resistance. In some embodiments, coating thickness may be tailored from 1-1000 pm depending on environment: 1-10 pm for minimum optical loss; 10-100 pm for harsher chemistries; up to 100-1000 pm for robust barriers. Representative application methods include coacervation or solvent evaporation for carbon-based polymers; spray coating (single or multi-pass) for controlled build; hot-melt encapsulation for epoxy-class systems; and interfacial polymerization, layer-by-layer (LbL) assembly, or atom transfer radical polymerization (ATRP)Attorney Docket No.: 064752-501001 WO for silicon- or hybrid-polymer coatings when nanometric control and high uniformity are required.[000195] In some embodiments, a photoreagent layer is applied directly over the barrier to convert incident photons into reactive species. Examples include TiCh (UV-A / near-UV; -350- 370 nm), ZnO (UV / blue), sulfite-bearing chemistries (^250-300 nm for hydrated electron generation), and indole-3 -acetic acid (-250-320 nm). In some embodiments, target thickness can be 10-100 nm to balance absorption with charge transport. Deposition methods can include dip coating, spray coating, spin coating, and sol-gel processes; post-deposition cures or photofixation steps can be used to stabilize the layer.[000196] In some embodiments, light-responsive layers provide reversible adsorption or dipole induction to stage reactants near the photoreagent. Illustrative chemistries include spiropyran / merocyanine systems (activated near -365 nm, reset near -550 nm) and azobenzene (UV-induced cis switching with thermal relaxation). These layers can be deposited as thin films over or adjacent to the photoreagent where reversible capture / release benefits selectivity or cleanup cycles.[000197] In some embodiments, an adsorbent porous layer overlies the photoreagent to preconcentrate reactants at the reactive interface. Representative materials include octadecylsilane (lipophilic), silica gels (hydrophilic, high surface area), and zeolites (molecular sieving). Thickness guidelines include: octadecylsilane -50-150 pm; silica gel -100-250 pm; zeolites -100-200 pm. Example pore structures include: octadecylsilane-coated porous media with 6-15 nm pores at -40-70% porosity; silica gels 2-50 nm, -60-80% porosity; zeolites -0.3-1.2 nm, -20-40% porosity. Silanization (e.g., using alkoxysilanes) may be used to covalently couple adsorbents to silica-primed barriers.[000198] In some embodiments where LED emitters are inserted directly into isoflux cells (e.g., through transparent windows or pocketed features), the same protective / photoreagent / adsorbent stack can be applied to the LED faces or windows rather than (or in addition to) the cell wall. In some embodiments, coating sequences respect optical planarity and maintain 8 by using calibrated hard-stops or gauge fixtures during cure. In some embodiments, where silane chemistry is used, LED-facing optics may receive a thin silica priming layer to enable robust coupling of photoreagents and adsorbents without degrading transmission.Attorney Docket No.: 064752-501001 WO[000199] In some embodiments, coating acceptance includes: spectral transmission and haze at the operating band; thickness verification (ellipsometry for 10-100 nm films; stylus / optical profilometry for 1-1000 m films); adhesion (cross-hatch / tape or pull-off on witness coupons); and chemical soak per target fluids. For adsorbent-over-photoreagent stacks, performance can be verified by areal adsorption capacity and by reaction-rate enhancement versus uncoated controls at identical irradiance.[000200] Example coating stacks (illustrative): (a) UV-AOP cell wall: SLA-clear wall —> silica primer — 20-60 nm TiCh sol-gel —> 150 pm porous silica (10-20 nm pores) silanized with octadecylsilane, (b) UV-ARP cell wall: COC wall —> fluoropolymer barrier 10-30 pm (spray) — > sulfite-bearing photolayer (spin, 20-80 nm) —> optional thin hydrophilic overcoat to manage wetting, (c) Visible-band bioreactor wall: PMMA / PC barrier 10-100 pm (spray / dip) — tailored adsorbent (silica or ODS) 100-200 pm. In some embodiments, no photocatalyst is used when purely photobiology is desired.[000201] In some embodiments, coatings are applied to individual shells prior to chain assembly to ensure coverage into fillets and interconnects; masked features protect sealing lands. Reconditioning may include solvent strip of the adsorbent layer while retaining the barrier, followed by re-application of the adsorbent and, if needed, refresh of the photoreagent thin film.Applications[000202] In some embodiments, the reactor is configured for photochemical operations across UV-visible bands (-365-700 nm and UV-C where applicable). Iso-flux cell chains, manifolded bundles, and LED options (2-, 4-, or wireless 6-LED) can enable use in liquid, gas-liquid, or slurry systems with tunable residence time and pressure drop.[000203] In some embodiments, the reactor performs direct UV photolysis and advanced oxidation / reduction processes, including photosensitized singlet-oxygen (AlOa) generation for destruction of trace organics. Examples include UV disinfection, H2O2 / UV-AOP, TiCh or ZnO photocatalysis, andA1 Ch-driven oxidation of pharmaceuticals and dyes in reuse polishing and point-of-use systems.[000204] In some embodiments, the reactor runs photoredox reductions / oxidations (Ir / Ru complexes, organic dyes, semiconductors), decarboxylative couplings, reductive dehalogenations, [2+2] cycloadditions, and photosensitized oxygenations via 1C>2. IllustrativeAttorney Docket No.: 064752-501001 WOA102 reactions include ene oxidations of alkenes to allylic hydroperoxides, [4+2] cycloadditions with 1,3-dienes to endoperoxides (e.g., anthracene, furans), and selective oxidations of sulfides to sulfoxides — useful for API / intermediate manufacture and route intensification.[000205] In some embodiments, the reactor serves as a photobioreactor for algae / cy anobacteria (e.g., blue / red bands for growth and product formation), photobiocatalysis(flavin / photoenzymes), and photoinduced gene expression. When desired, controlled photosensitizer dosing produces low levels ofA1O2for targeted oxidative transformations in cell- free or enzyme systems.[000206] In some embodiments, the reactor supports gas-liquid photochemistry for odor / VOC abatement and synthesis of specialty monomers by photopolymerization or halogenation (where permitted). Slurry-compatible cells enable immobilized photocatalysts or recyclable photosensitizers forA1O2processes in continuous flow.[000207] In some embodiments, the reactor executes light-driven steps in GMP or pilot settings, including photoredox C-N / C-C formations, oxidative rearrangements, and singletoxygenations (ene / endoperoxide pathways) that benefit from tight dose control and scalable, parallelized chains.[000208] In some embodiments, the reactor is used for UV-based sterilization, pigment stabilization, controlled photoisomerizations for fragrance / flavor, and mild 1O2treatments for functional ingredient modification. For each application, wavelength, irradiance, residence time, and allowable pressure drop are selected per the design methods herein; cells, chains, and LED configurations are chosen to meet target dose while observing manufacturing limits (e.g., <100 cm chain length per SLA build). These examples illustrate breadth of use and are not limiting.[000209] FIG. 12 is a block diagram 1200 of a computing system 1210 suitable for use in implementing the computerized components described herein (e.g., the workflow 1000 shown in FIG. 10). In broad overview, the computing system 1210 includes at least one processor 1250 for performing actions in accordance with instructions, and one or more memory devices 1260 and / or 1270 for storing instructions and data. The illustrated example computing system 1210 includes one or more processors 1250 in communication, via a bus 1215, with memory 1270 and with at least one network interface controller 1220 with a network interface 1225 for connecting to external devices 1230, e.g., a computing device. The one or more processors 1250 are also in communication, via the bus 1215, with each other and with any I / O devices at one or more I / OAttorney Docket No.: 064752-501001 WO interfaces 1240, and any other devices 1280. The processor 1250 illustrated incorporates, or is directly connected to, cache memory 1260. Generally, a processor will execute instructions received from memory. In some embodiments, the computing system 1210 can be configured within a cloud computing environment, a virtual or containerized computing environment, and / or a web-based microservices environment.[000210] In more detail, the processor 1250 can be any logic circuitry that processes instructions, e.g., instructions fetched from the memory 1270 or cache 1260. In many embodiments, the processor 1250 is an embedded processor, a microprocessor unit or special purpose processor. The computing system 1210 can be based on any processor, e.g., suitable digital signal processor (DSP), or set of processors, capable of operating as described herein. In some embodiments, the processor 1250 can be a single core or multi-core processor. In some embodiments, the processor 1250 can be composed of multiple processors.[000211] The memory 1270 can be any device suitable for storing computer readable data. The memory 1270 can be a device with fixed storage or a device for reading removable storage media. Examples include all forms of non-volatile memory, media and memory devices, semiconductor memory devices (e.g., EPROM, EEPROM, SDRAM, flash memory devices, and all types of solid-state memory), magnetic disks, and magneto optical disks. A computing device 1210 can have any number of memory devices 1270.[000212] The cache memory 1260 is generally a form of high-speed computer memory placed in close proximity to the processor 1250 for fast read / write times. In some implementations, the cache memory 1260 is part of, or on the same chip as, the processor 1250.[000213] The network interface controller 1220 manages data exchanges via the network interface 1225. The network interface controller 1220 handles the physical, media access control, and data link layers of the Open Systems Interconnect (OSI) model for network communication. In some implementations, some of the network interface controller’s tasks are handled by the processor 1250. In some implementations, the network interface controller 1220 is part of the processor 1250. In some implementations, a computing device 1210 has multiple network interface controllers 1220. In some implementations, the network interface 1225 is a connection point for a physical network link, e.g., an RJ 45 connector. In some implementations, the network interface controller 1220 supports wireless network connections and an interface port 1225 is a wireless Bluetooth transceiver. Generally, a computing device 1210 exchanges dataAttorney Docket No.: 064752-501001 WO with other network devices 1230, such as computing device 1230, via physical or wireless links to a network interface 1225. In some implementations, the network interface controller 1220 implements a network protocol such as LTE, TCP / IP Ethernet, IEEE 802.11, IEEE 802.16, Bluetooth, or the like.[000214] The other computing devices 1230 are connected to the computing device 1210 via a network interface port 1225. The other computing device 1230 can be a peer computing device, a network device, a server, or any other computing device with network functionality. In some embodiments, the computing device 1230 can be a network device such as a hub, a bridge, a switch, or a router, connecting the computing device 1210 to a data network such as the Internet.[000215] In some uses, the I / O interface 1240 supports an input device and / or an output device (not shown). In some uses, the input device and the output device are integrated into the same hardware, e.g., as in a touch screen. In some uses, such as in a server context, there is no I / O interface 1240 or the I / O interface 1240 is not used. In some uses, additional other components 1280 are in communication with the computer system 1210, e.g., external devices connected via a universal serial bus (USB).[000216] The other devices 1280 can include an I / O interface 1240, external serial device ports, and any additional co-processors. For example, a computing system 1210 can include an interface (e.g., a universal serial bus (USB) interface, or the like) for connecting input devices (e.g., a keyboard, microphone, mouse, or other pointing device), output devices (e.g., video display, speaker, refreshable Braille terminal, or printer), or additional memory devices (e.g., portable flash drive or external media drive). In some implementations an I / O device is incorporated into the computing system 1210, e.g., a touch screen on a tablet device. In some implementations, a computing device 1210 includes an additional device 1280 such as a coprocessor, e.g., a math co-processor that can assist the processor 1250 with high precision or complex calculations.[000217] While this disclosure contains many specific implementation details, these should not be construed as limitations on the scope of what may be claimed, but rather as descriptions of features specific to particular implementations of particular inventions. Certain features that are described in this disclosure in the context of separate implementations can also be implemented in combination in a single implementation. Conversely, various features that are described in the context of a single implementation can also be implemented in multiple implementationsAttorney Docket No.: 064752-501001 WO separately or in any suitable sub-combination. Moreover, although features may be described above as acting in certain combinations and even initially claimed as such, one or more features from a claimed combination can in some cases be excised from the combination, and the claimed combination may be directed to a sub-combination or variation of a sub-combination.[000218] Similarly, while operations are depicted in the drawings in a particular order, this should not be understood as requiring that such operations be performed in the particular order shown or in sequential order, or that all illustrated operations be performed, to achieve desirable results. Moreover, the separation of various system components in the implementations described above should not be understood as requiring such separation in all implementations, and it should be understood that the described components and systems can generally be integrated together in a single product or packaged into multiple products.[000219] Thus, particular implementations of the subject matter have been described. Other implementations are within the scope of the following claims. In some cases, the actions recited in the claims can be performed in a different order and still achieve desirable results. In addition, the processes depicted in the accompanying figures do not necessarily require the particular order shown, or sequential order, to achieve desirable results.

Claims

Attorney Docket No.: 064752-501001 WOCLAIMS1. A flow-cell device for a photochemical reactor, comprising: a body defining an interior volume bounded by an inner wall, wherein: the inner wall is shaped to substantially coincide with an iso-flux boundary at a target photon-flux threshold ( *), the iso-flux boundary being the locus of points at which a photon-flux density, as evaluated by an attenuation model accounting for absorption (a) at an operating wavelength and emitter angular emission, equals <D* under a design illumination layout and an optical standoff distance (8) between emitters and the inner wall; the design illumination layout comprises at least one of (i) a two-emitter layout comprising two emitters as an opposed pair on a common axis, (ii) a four-emitter layout comprising four emitters as two opposed pairs on two mutually orthogonal axes, and (iii) a six- emitter layout comprising six emitters as three opposed pairs on three mutually orthogonal axes; and the inner-wall shape is selected from a parametric family of shapes defined by a, C>*, 8, and the selected layout.

2. The device of claim 1 , wherein the body includes emitter-reference features comprising planar standoff rails or keyed pockets positioned to locate light-emitting faces at the standoff 8.

3. The device of claim 1 , wherein the layout is the two-emitter layout, an emitter axis defines a symmetry axis, and the inner wall exhibits axial symmetry about that axis and mirror symmetry across a midplane normal to the symmetry axis.

4. The device of claim 1 , wherein the layout is the four-emitter layout, two orthogonal emitter axes define symmetry axes, and the inner wall exhibits two orthogonal mirror planes and bilateral symmetry with respect to each axis.

5. The device of claim 1 , wherein the layout is the six-emitter layout, three mutually orthogonal emitter axes define symmetry axes, and the inner wall exhibits triaxial symmetry with respect to those axes.

6. The device of claim 1 , wherein a coincidence tolerance between the inner wall and the iso-flux boundary is a maximum normal offset less than or equal to 0.20 R_eff.Atorney Docket No.: 064752-501001 WO7. The device of claim 6, wherein emiters are located outside the weted volume, and the emiter-reference features set 8 relative to exterior surfaces such that the iso-flux boundary coincides with the inner wall within the coincidence tolerance.

8. The device of claim 1 , wherein the coefficient of variation of photon-flux density over {<D > O*} is less than or equal to 0.2 under the design illumination layout at standoff 8.

9. The device of claim 1 , wherein the maximum normal offset between the inner wall and the iso-flux boundary is less than or equal to 0.10 R_eff.

10. The device of claim 1 , wherein the maximum normal offset between the inner wall and the iso-flux boundary is less than or equal to 0.05 R_eff.

11. The device of claim 1 , wherein 0.1 R_eff < 8 < 0.5 R_eff.

12. The device of claim 1 , wherein the inner wall comprises smoothly joined surfaces whose minimum radius of curvature satisfies r min > O.l R efif.

13. The device of claim 1, wherein the parametric family of shapes is computed or generated by an optimization that minimizes a uniformity metric over { > *} while constraining 8 and the selected layout.

14. The device of claim 1 , wherein the inner wall is represented in manufacture by a superellipsoid (|x / a|Am + |y / b|Am + |z / c|Am)A(l / m)=l with parameters a,b,c,m chosen such that the maximum normal offset to the iso-flux boundary at <D* is < 0.20 R eff.

15. The device of claim 1 , wherein D_min is a minimum principal axis length of the flow cell, defined as the smallest principal diameter measured between opposing points on the inner wall along any illumination-defined principal axis, and the selected shape satisfies min(2a, 2b, 2c) > D_min, where a, b, c are principal half-spans along those axes.

16. The device of claim 6, wherein if a computed family member would yield a principal diameter < D_min, the selected member is the nearest member that meets D_min while maintaining the coincidence tolerance.

17. The device of claim 1, wherein the device is configured to mate with cylindrical interconnect capillaries of inner diameter d_c and length L_c satisfying d_c < 0.2 R_eff and L_c < 0.5 R_eff.Atorney Docket No.: 064752-501001 WO18. The device of claim 1, wherein emiter face centers lie on an emiter-reference surface S3 defined as the outer normal offset of the inner wall by 8, and emitter face normals are orthogonal to S3 within ±3°.

19. The device of claim 18, wherein emiter face normals deviate < 3° from the local normal of S3 and emiter-to-wall standoff deviates from 8 by less than or equal to 0.10 R_eff.

20. The device of claim 2, wherein the two emiter centers are colinear on the symmetry axis and approximately equidistant from a geometric center of the flow cell.

21. The device of claim 3, wherein emiter centers form two opposed pairs on mutually orthogonal axes, each center lies on S3, and pairwise distances along each axis are approximately equal.

22. The device of claim 4, wherein the six emiter centers form an octahedral set on three mutually orthogonal axes, each center lies on S3, and all six center-to-center distances from a cell center are equal within ±10%.

23. A photochemical reactor apparatus, comprising: a headplate with an array of through-ports arranged on a latice of lateral pitch (s); a plurality of flow cell chains, each chain comprising a sequence of flow cells connected in series by cylindrical interconnect capillaries of inner diameter (d_c) and length (L_c).

24. The apparatus of claim 23, wherein each flow cell comprises an inner wall shaped to substantially coincide with an iso-flux boundary at a target photon-flux threshold (O*) with a maximum normal-offset tolerance < 0.20 R_eff, the iso-flux boundary being evaluated under a design illumination layout comprising at least one of: (i) two emiters as an opposed pair on a common axis; (ii) four emiters as two opposed pairs on two mutually orthogonal axes; and (iii) six emiters as three opposed pairs on three mutually orthogonal axes, wherein an attenuation model is used to account for absorption (a) at the operating wavelength and emiter angular emission, with an optical standoff distance (8) between emiters and the inner wall.

25. The apparatus of claim 23, wherein inlet and outlet headers sealed to opposite faces of the headplate and formed as dished plenums in fluid communication with the through-ports.

26. The apparatus of claim 23, further comprising a plurality of LED emiters configured to illuminate flow cells of the sequence of flow cells, wherein the plurality of LED emiters areAttorney Docket No.: 064752-501001 WO arranged according to a design illumination layout, wherein the design illumination layout comprises at least one of: (i) a two-emitter layout comprising two planar LED plates positioned on opposite sides of a flow-cell chain and registered by standoff spacers to set 8; (ii) a four- emitter layout comprising LED strips mounted on four faces of a rectangular LED column positioned between adjacent chains to present two opposed pairs around each neighboring cell; and (iii) a six-emitter layout comprising a unified chain in which cube-LED positions and flowcell positions are mechanically integrated at a repetition interval pitch (p) and provided in flow- first and cube-first types installed in an alternating, laterally-staggered pattern so that each flow cell is surrounded by six cube-LED positions and each cube-LED position is surrounded by six flow-cell positions, each at standoff 8 within a maximum normal-offset tolerance < 0.20 R_efif; wherein for each flow cell chain the interconnect capillaries satisfy d_c < 0.2 R_eff and L_c < 0.5 R_eff.

27. The apparatus of claim 26, wherein headplate registration features comprising sockets or rails set the lateral pitch (s) and a stagger offset to maintain 8 within the maximum normal-offset tolerance across the bundle.

28. The apparatus of claim 26, wherein, for the two-emitter layout, each planar LED plate comprises a two-dimensional LED array on a thermally conductive substrate and includes manifolded microchannels or a serpentine liquid passage with inlet and outlet ports routed outside the process volume.

29. The apparatus of claim 26, wherein, for the two-emitter layout, adjacent chains share back-to-back LED plates mounted to a common carrier that includes cooling passages or sealed cooling corridors hydraulically isolated from the process fluid.

30. The apparatus of claim 26, wherein, for the four-emitter layout face-to-chain spacing equals 8 on each side so the iso-flux boundary remains coincident with the inner wall, and the columns and chains are arranged on a square checkerboard lattice such that each chain is bounded by four LED columns and each column serves four chains.

31. The apparatus of claim 26, wherein, for the four-emitter layout, each LED column includes a thermally conductive core with cooling passages or sealed cooling corridors hydraulically isolated from the process fluid.Attorney Docket No.: 064752-501001 WO32. The apparatus of claim 26, wherein, for the four-emitter layout, the LED-strip emitter centers are indexed to cell midplanes at the pitch (p) within ±0.2 p, and strip planes are parallel to a chain axis within ±3°.

33. The apparatus of claim 26, wherein, for the six-emitter layout, the unified chain is a wired chain and conductors for the cube-LEDs are routed along the chain to a feedthrough on a dry side.

34. The apparatus of claim 26, wherein, for the six-emitter layout, the unified chain uses wireless cube-LED assemblies that are inductively powered by at least one transmitter coil, and the assemblies are positioned so that emitter face centers lie on an emitter-reference surface S3 defined as the outer normal offset of the inner wall by 8, with face normals orthogonal to S3 within ±3°, and spacing set to 8 within the maximum normal-offset tolerance.

35. The apparatus of claim 26, wherein centers of the six lateral cube-LED positions around each flow cell define a regular hexagon in plan with edge-length uniformity within ±10% across the bundle.

36. A wireless cube LED assembly for illuminating photochemical flow cells, comprising: a polyhedral body having six faces arranged as three mutually orthogonal opposed pairs; six LED emitters, one on each face of the six faces; a receiver coil disposed within the body; a rectifier coupled to the receiver coil; a constant-current driver coupled to the six LED emitters; and alignment features configured to set a nominal optical standoff (8) and orthogonal orientation relative to a mating surface, the assembly being configured to receive power inductively from a transmitter coil at a transmitter frequency (f_T).

37. The assembly of claim 36, wherein an edge length of the body is 5-20 mm.

38. The assembly of claim 36, wherein an emission band of the LED emitters is 200-700 nm with a per-face radiant flux of 10-1000 mW at 25 °C.

39. The assembly of claim 36, wherein an inductance of the receiver coil is 0.1-10 mH and the transmitter frequency is 100 Hz-100 kHzAttorney Docket No.: 064752-501001 WO40. The assembly of claim 36, wherein the faces are fabricated on FR-4 or aluminum-core PCB and the body is filled or coated with potting selected from silicone or fluorosilicone and optionally parylene.

41. The assembly of claim 36, wherein the transmitter coil is coupled to a controller configured to sweep the transmitter frequency across a band with assemblies installed in a reactor and to select an operating frequency (f_T*) that maximizes received power or minimizes primary current for a specified secondary load.

42. The assembly of claim 32, wherein the assembly is configured to be inserted into keyed pockets of a flow-cell chain at planes separated by a pitch (p), and when installed the emitter face centers lie on an emitter-reference surface S3 defined as the outer normal offset of the inner wall by 8, with face normals orthogonal to S3 within ±1°.

43. The apparatus of claim 23, wherein a chain axis is defined by centers of successive flow cells in a chain, and an LED array centerline of an emitter plate, strip column, or cube-LED spine is parallel to the chain axis within ±1° and has a maximum lateral offset Al < 0.20 R_eff relative to the chain axis along at least 80% of a chain length.

44. The apparatus of claim 26, wherein, for the two-emitter layout, each planar LED plate has an emitter line that is coincident with the chain axis within Al < 0.1 R eff, and the plate normals are arranged symmetrically about a chain midplane within ±3°.

45. The apparatus of claim 26, wherein, for the four-emitter layout, each LED strip column has a column axis registered to the adjacent cell-chain axis with Al < O.lO R eff and yaw / pitch < 3°.

46. The apparatus of claim 24, wherein standoff uniformity satisfies |8(s) - 8| < 0.1 R_eff along any emitter face spanning a bulb, within the maximum normal-offset tolerance.

47. The apparatus of claim 46, wherein bulb-to-neck geometry satisfies a diameter ratio D bulb / d_c > 2.0 for each cell, where D bulb is the maximum inner diameter of the bulb measured on a plane through the cell center and d_c is the interconnect inner diameter.

48. The apparatus of claim 23, wherein emitter indexing to the cell pitch p satisfies a midplane registration error < 0.1 p for the positions of emitter centers relative to corresponding cell midplanes.Attorney Docket No.: 064752-501001 WO49. The apparatus of claim 26, wherein, for the six-emitter layout, the six lateral cube-LED centers around each flow cell lie on a ring of radius R_hex in plan view, with |R_hex - (8 + t_wall)| < O.lO R eff and edge-length uniformity within ±10%, where t_wall is a local wall thickness at the bulb equator.

50. The apparatus of claim 26, wherein, for the six-emitter layout, the axial cube-LED centers are located on planes separated by the pitch p and are concentric with the cell center within Al < 0.1 R_eff, and face normals are orthogonal to S8 within ±3°.

51. The device of claim 1 , wherein the inner wall carries a protective barrier coating that is chemically resistant and optically transmissive at the operating wavelength, the barrier being selected from fluoropolymers, cyclic olefin copolymers (COC / COGS), polyacrylates / epoxies, PMMA, or polycarbonate.

52. The device of claim 51 , further comprising a photoreagent layer over the barrier configured to convert incident photons into reactive species, the photoreagent being selected from TiO2, ZnO, sulfite-bearing chemistries, or indole-3 -acetic acid, with a target thickness 10- 100 nm.

53. The device of claim 51 or 52, further comprising an adsorbent porous layer overlying the barrier or the photoreagent to pre-concentrate reactants at the reactive interface, the adsorbent being selected from silica gels, zeolites, or octadecylsilane-treated porous media.

54. The device of claim 1 , wherein the inner wall carries a reversibly switchable dipolar surface layer configured to modulate adsorption / desorption for chromatography or capture / release operations, the layer including spiropyran / merocyanine or azobenzene chemistries operable by optical stimulus.

55. The device of any of claims 51-54, wherein one or more coatings are applied by spray, dip, spin, sol-gel, interfacial polymerization, or layer-by-layer assembly, and the resulting stack maintains a coincidence tolerance of less than or equal to 0.20 R_eff.

56. The device of claim 1 , wherein a dry-side reflective layer is disposed behind a transparent inner wall region to improve photon-flux uniformity, the reflective layer being specular (protected Al or Ag) or dielectric stack or diffuse PTFE-class, and arranged so that the coefficient of variation over {<D > <D*} is < 0.2.Attorney Docket No.: 064752-501001 WO57. The device of claim 1 , wherein windows are bonded into flush pockets of the inner wall, the windows having anti-reflection coatings on an air side and sealed by welded frames or compatible adhesives, while not shifting the iso-flux boundary by more than 0.20 R eff.

58. The apparatus of claim 24, comprising a diffuser optic on the emitter side selected from micro-structured, holographic, molded, or ground-glass diffusers having a scattering half-angle sufficient to reduce the coefficient of variation over {O > <!>*} to < 0.2 while maintaining 5 within the maximum normal-offset tolerance.

59. The apparatus of claim 26, wherein, for the two-emitter layout and the four-emitter layout, planar LED plates or LED strip columns further include reflective baffles or liners positioned to return spill light toward the inner wall.

60. The device of claim 1 , wherein emitters are inserted directly into keyed pockets that open to the wetted volume and are sealed by a transparent encapsulant or window, the optical plane of the encapsulant / window being co-planar with the inner wall within the tolerance of less than or equal to 0.20 R eff, and 8 being gauged to said optical plane.

61. The device of claim 60, wherein the exposed window or encapsulant face carries one or more of the coatings of claims 51-56.

62. The apparatus of claim 24, wherein, alternatively, emitters remain outside the wetted volume behind windows bonded to the cell, the windows having anti-reflection coatings and the face-to-wall spacing set to 8 so that the iso-flux boundary remains coincident with the inner wall within the maximum normal-offset tolerance.

Images